A Probabilistic Forecasting of Hurricane Induced Storm Surge

Abstract:

Tropical cyclones (TCs) are a hazard to life and property and a prominent element of the global climate system, thus understanding and predicting their intensity, frequency and possible impacts is of paramount importance to the society. In coastal cities, the determination of the levels of defense against flooding is based on the best estimate of the probability of the hurricane induced storm surge. This certainly involves estimating the probability that a hurricane will pass within some given radius of the point in question, yet there are too few such events in the historical record to infer robust probabilities from the data alone, casting at least some doubt on the historical estimates. Recent research efforts have indicated that one partial solution to this problem is to generate a very large number of synthetic hurricanes, whose most important statistical properties conform to those of historical events. This work presents a cascade modelling approach for the prediction of storm surges in Mexico. The approach involves the numerical simulations of wind, waves and hydrodynamics induced by 3100 synthetic hurricane tracks, in comparison to that determined from the historical events. For this, wind fields are used as forcing in a third generation wave model and a 2D hydrodynamic model, enabling the characterization of hurricane conditions from the synthetic events. For several coastal locations on Mexico, we present a comparison of the estimates with this approach against those directly derived from historical hurricane data. The results suggest that such modelling approach opens the door towards a robust dynamical forecast of TC induced storm surges.